Whither the Bicarbonate Era

To the Editor:

For metabolic acidosis, beyond treating the underlying cause, correcting hypoxemia, and establishing good perfusion, sodium bicarbonate is often given at variable arbitrary thresholds of depressed blood pH. Recently, Zanella and colleagues (1) employed extracorporeal removal of chloride by electrodialysis in healthy pigs made acidemic by either lactic acid infusion or hypoventilation (CO₂ retention). By physically drawing off chloride and establishing a local separation of charge, blood electroneutrality at the membrane is immediately reestablished by the hydrolysis of water to yield a hydroxyl ion that instantly combines with CO₂ to form bicarbonate. The authors show the feasibility of quantitatively increasing bicarbonate in this fashion for both forms of acidosis without the associated and unwanted hypernatremia and volume loading that can occur with intravenous sodium bicarbonate administration. The accompanying editorialists (2) proclaim the postbicarbonate era with this study that illustrates a major tenet of the Stewart approach to acid–base chemistry and its superiority over other approaches to understanding acid–base physiology and pathophysiology. In Stewart’s paradigm, H⁺, OH⁻, HCO₃⁻, and CO₃²⁻ are relegated to the status of dependent variables; that is, they can only be formed from the differential movements and exchanges of independent strong ions (Na⁺, K⁺, and Cl⁻) that disturb electroneutrality, which is immediately corrected by the hydrolysis of water and reaction with CO₂. Although the heuristics of the Stewart analysis are valid, I remain unconvinced by the claim that what the mathematics of this approach reveal demands that physiology must follow these rules and conclusions. The assumption that only strong ions and their differential movement from one space to another alters H⁺ and HCO₃⁻ concentrations because the math is consistent with it goes against all we know about numerous cell-membrane transporters that use H⁺, HCO₃⁻, or CO₃²⁻ as coions or counterions with Na⁺, K⁺, and Cl⁻ (3). There has been no identification of a Na⁺/Cl⁻ antiporter or an electrogenic Na⁺-Cl cotransporter that will alter local electroneutrality to create or consume the supposed dependent variables. Simply because one can electrodialyze chloride by brute force, as Zanella and colleagues (4) report, does not mean it happens in vivo at the microscopic level. Clinical adoption of the Stewart approach has shown no superiority to conventional approaches (4–6), and the measurement of all the strong ions repeatedly is wasteful and costly, risks measurement errors that beget more testing, contributes to anemia in many patients, contributes to more transfusions, and is very difficult to teach to trainees and even seasoned clinicians.